AiCostCompare

Compare models

Pick up to 4 models for a side-by-side spec sheet. Hover any dotted metric name for what the test measures. Best value in each row is highlighted.

Claude Haiku 4.5MiMo-V2.5
Claude Haiku 4.5

Anthropic

MiMo-V2.5

Xiaomi

Pricing
Input $/1M$1.00$0.14
Output $/1M$5.00$0.28
Blended $/1M (3:1)$2.00$0.175
Cache read $/1M$0.10$0.0028
Cache write $/1M$1.25
Quality indices
Intelligence Index29.637.2
Coding Index43.956.8
Math Index
Agentic Index16.423.7
Speed & latency
Output speed (tok/s)63
Time to first token3.68s
Time to first answer token35.63s
Benchmarks (Artificial Analysis)
GPQA Diamond84.9%
Humanity's Last Exam25.2%
SciCode43.1%
IFBench67.1%
AA-LCR (Long Context Reasoning)62.7%
Terminal-Bench Hard41.7%
Terminal-Bench 2.163.7%
τ²-Bench (Telecom)90.6%
τ³-Bench Banking6.6%
Design Arena (Elo)
Website1,1491,288
UI components1,1421,294
Data viz1,1601,279
SVG1,0721,218
3D1,1311,273
Game dev1,1551,291
Code categories1,1481,286
ASCII art1,1821,175
Specs
Context window200K1.05M
Max output tokens64K131K
Input modalitiestext, image, filetext, audio, image, video
TokenizerClaudeOther